{
  "id": 128572,
  "title": "Why is a horizontally flipped image treated as a new ID",
  "url": "/competitions/humpback-whale-identification/discussion/128572",
  "author_name": "",
  "post_date": "2020-02-01T12:23:00.746619300Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I am a novice, I do not quite understand why the master of the horizontal flip pictures as a new category, the addition of new classes isn't more difficult to learn?</p>",
  "messages": [
    {
      "id": "734405",
      "postDate": "02/01/2020 12:23:00",
      "content": "<p>I am a novice, I do not quite understand why the master of the horizontal flip pictures as a new category, the addition of new classes isn't more difficult to learn?</p>",
      "rawMarkdown": "I am a novice, I do not quite understand why the master of the horizontal flip pictures as a new category, the addition of new classes isn't more difficult to learn?",
      "votes": null
    },
    {
      "id": "734692",
      "postDate": "02/01/2020 20:53:10",
      "content": "<p>Getting our data trained from every possible view can help our network learn many things from same image so we normalize the image and try passing it with variance like flipping,mirroring,rotating, etc</p>",
      "rawMarkdown": "Getting our data trained from every possible view can help our network learn many things from same image so we normalize the image and try passing it with variance like flipping,mirroring,rotating, etc",
      "votes": null
    },
    {
      "id": "740552",
      "postDate": "02/09/2020 14:13:10",
      "content": "<p>Thanks,your answer help me to understand this thing</p>",
      "rawMarkdown": "Thanks,your answer help me to understand this thing",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 734692,
      "author_name": "chintanchitroda",
      "author_url": "",
      "post_date": "02/01/2020 20:53:10",
      "content": "<p>Getting our data trained from every possible view can help our network learn many things from same image so we normalize the image and try passing it with variance like flipping,mirroring,rotating, etc</p>",
      "votes": null,
      "replies": [
        {
          "id": 740552,
          "author_name": "ssfailearning",
          "author_url": "",
          "post_date": "02/09/2020 14:13:10",
          "content": "<p>Thanks,your answer help me to understand this thing</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "734405": "I am a novice, I do not quite understand why the master of the horizontal flip pictures as a new category, the addition of new classes isn't more difficult to learn?",
    "734692": "Getting our data trained from every possible view can help our network learn many things from same image so we normalize the image and try passing it with variance like flipping,mirroring,rotating, etc",
    "740552": "Thanks,your answer help me to understand this thing"
  },
  "source": "meta"
}